►
From YouTube: Sarah Polan Scaling the 4th Industrial Revolution
Description
With the Cloud Native landscape rapidly expanding and Kubernetes being used in over 50% of organizations, we've never had more opportunities for innovation and disruption, nor has the expectation for success been so demanding. With executive demands for multi-cloud interoperability, implementing cutting edge technology, and visible cost reduction across all departments, how can we ensure continued innovation? How do we manage that at scale across diverse technology and cultures? We'll dive into the pillars required for adopting developer centric platforms across organizations at scale in order to facilitate developer velocity.
A
All
right,
first
of
all,
thank
you
so
much
for
joining
me
this
morning.
I
am
absolutely
floored
by
the
number
of
people
who
are
here
to
hear
about
the
fourth
Industrial
Revolution
at
9
30.
In
the
morning
my
name
is
Sarah
I'm,
the
field
CTO
for
Mia
with
hashicorp.
A
For
those
of
you
who
don't
know,
hashicorp
we're
Community
oriented
software
provider,
we
do
terraform
Vault
and
our
goal
is
really
to
provide
secure
operations
and
operation
provisioning
across
the
board,
regardless
of
what
kind
of
environment
you're
in
so
whether
that's
kubernetes,
on-prem
public
cloud
or
a
mix
thereof.
A
So
as
field
CTO,
one
of
the
things
that
I
get
to
do
and
I
think
I
probably
have
one
of
the
coolest
jobs
out.
There
is
I,
spend
a
lot
of
time
with
the
Fortune
500
and
with
the
Fortune
500
I
help
advise
them
on
how
to
get
the
most
out
of
their
Cloud
operating
models
little
tweaks
and
things
that
they
can
do
to
really
make
sure
that
that
operability
is
there.
A
I
put
a
lot
of
time
and
effort
into
figuring
out
how
we
can
kind
of
start
to
unify
those
two
things,
because,
as
we
see
this
Market
evolve
and
as
we
see
what's
happening
just
generally
speaking
in
the
world,
it's
no
longer
possible
for
the
security
people
to
say
no
and
it's
no
longer
possible
for
the
developers
and
the
engineers
to
say.
Yeah
security
is
not
our
problem,
so
it's
really
about.
A
A
A
A
What
I
made
the
mistake
of
doing
was
actually
conflating
those
first
two
industrial
revolutions,
so
thinking
that
some
of
this
mechanization
steam
power
actually
was
part
of
that
assembly
line
and
moving
a
little
bit
more
into
automotive
and
things
like
that.
Well,
it
turns
out
these
two
things
are
very
separate.
A
Third
Industrial
Revolution.
Perhaps
my
favorite
and
probably
then
dates
me-
is
about
computers,
technology
and
again,
automation
within
factories.
So
things
like
eniac,
aatrax,
VHS
Nintendo,
all
those
good
things,
but
also
then
I
think
towards
the
movie
Willy
Wonka,
and
you
know
the
dad
who's
screwing,
the
the
toothpaste
caps
gets
replaced
by
Machinery
because
the
Machinery
is
faster.
A
All
of
these
things
start
fuzzing
that
line
between
digital
and
physical
and
bringing
that
together
and
as
we
see
that
happen
what's
happening,
is
that
we're
changing
business
models,
we're
changing
adoption
models,
and
so
that's
also
having
an
impact
on
how
we
scale
people
how
we
manage
things
like
smart
devices
because
they
have
to
be
connected,
but
at
the
same
time
they're
tangible
assets
that
we
need
to
manage
now.
The
last
thing
I
want
to
mention
about
this
is,
if
you
look
at
all
of
these,
each
one
is
built
on
the
last.
A
So
next
random
topic
I
want
to
throw
at
you
is
our
friend,
Occam
and
Occam's
razor.
So
William
of
Occam
is
an
English
logician.
Also
Franciscan
fryer
and
he
says
entities
should
not
be
multiplied
unnecessarily.
A
So
what
does
that
mean?
Well,
basically,
it's
what
we
consider
to
be
a
practical
guideline:
it's
not
a
law,
it's
not
natural
law,
but
it's
something
that
we
can
base
yeah,
how
we
develop
and
how
we
how
we
move
forward.
Essentially
we
see
it
a
lot
happen
in
science
and
technology
and
it's
basically
yeah
if
there
are
multiple
explanations
of
a
hypothesis.
A
So
it's
important
that
we
note,
though,
that
that's
within
a
particular
context,
so
if
I'm
operating
within
that
context,
I
need
to
be
able
to
make
a
decision
based
off
of
the
fewest
assumptions
possible
and
take
the
fewest
steps
forward,
while
I'm
taking
into
account
my
environment
and
that's
most
likely
going
to
lead
me
to
the
best
solution,
not
always,
but
mostly
so
advice
that
I
don't
take
very
regularly.
Is
that
probably
avoid
over-analyzing
situations?
A
A
This
paper
was
actually
initially
rejected
by
the
Harvard
Business
Review,
because
it
said
he
didn't
prove
his
thesis.
This
is
something
that
today
across
software
engineering,
we
hold
to
be
relatively
true
and
we
actually
call
it
a
law
now,
I'm
sure.
A
lot
of
you
have
heard
of
Conway's
law,
for
those
of
you
who
haven't
Conway's
law
says
that
the
architecture
and
design
of
a
system
will
directly
correlate
to
the
communication
structure
within
an
organization.
A
A
All
right,
so
your
next
next
Theory
lesson
for
today,
Willow
Run,
now
Willow
Run.
For
those
of
you
who
don't
know
your
American
history
shocking
was
actually
a
factory,
a
manufacturing
complex
that
was
located
in
Michigan
and
it
was
run
by
the
Ford
automotive
company
to
create
bombers
during
World
War
II.
A
So
the
problem
was
that
at
the
end
of
1939
there
were
approximately
1200
bombers
left,
which
was
not
enough,
and
this
Fleet
had
undergone
immense
shrinkage
due
to
Blitzkrieg,
basically
so
Total
War
velt
at
that
time
said
that's
great.
We
have
1200
bombers,
I
need
50
000
by
the
end
of
the
year,
oh,
and
by
the
way,
I
need
50,
000
the
year
after
as
well.
A
So
then
what
Ford
did
was
he
came
in.
He
overhauled
the
entire
system.
It
wasn't
perfect
at
first,
there
were
still
some
major
issues,
but
it
was
an
iterative
process
and
basically
by
using
common,
reusable
parts,
specialist
functions
assembly
lines.
Let's
see
where
I'm
going
here
the
plant
was
able
to
reduce
the
speed
of
creation
of
a
bomber
to
63
minutes,
that's
down
from
1752
minutes
for
a
single
bomber
earlier
in
the
year.
A
Last
military
concept
of
the
day,
the
ooda
loop,
so
the
ooda
loop
is
a
concept
that
was
created
by
excuse
me.
First
name
forget
it
anyway.
Colonel
Boyd
and
Colonel
Boyd
came
up
with
this
idea
that,
in
order
to
train
a
pilot
and
to
have
them
make
the
best
decision,
they
need
to
have
this
Loop
in
place.
A
So
it
stands
for
observe,
Orient,
decide
and
act.
It's
a
four-step
process
and
it's
something
that
really
guides
particularly
air,
combat,
Pilots
and
the
reason
that
Boyd
came
up
with
this
was
when
he
was
in
the
Korean
War.
He
noticed
that
the
best
Pilots
weren't,
the
ones
that
had
the
best
technology
they
weren't
the
ones
who
had
the
best
machinery.
A
So
as
a
result,
what
they
did
is
they
started
training
the
Air
Force
and
they
started
training
the
Air
Force
to
observe
Orient,
decide
and
act,
because
what
we've
seen
is
that
in
competitive
situations,
the
faster
you
can
complete
this
Loop.
If
you
can
complete
it
faster
than
your
competitor,
you
actually
have
a
higher
rate
of
success.
A
So
now
it's
still
used
in
military.
It's
still
used
in
military
strategy.
It's
also
something
that
we've
actually
adopted
into
the
business
world,
particularly
in
highly
competitive
situations
and
I.
Think
this
is
something
that
we
see,
particularly
with
the
large
csps,
for
example,
they're
constantly
competing
against
each
other
iterating
coming
up
with
different
Services
products,
so
that
they
can
be
first
on
the
market
so
that
they
can
have
the
larger
market
segment.
A
So
let's
talk
a
little
bit
about
it
now
when
I
work
with
different
leaders
in
the
industry,
what
I
really
see
is
that
we
have
three
common
goals
and
within
these
three
common
goals
and
these
strategies,
some
are
more
successful
than
others.
If
I'm
going
to
be
really
honest,
we
see
all
ends
of
the
spectrum.
We
have
certain
companies
that
are
very
on
top
of
it,
leading
Cutting,
Edge
technology
and
fulfill
all
three
of
these.
We
have
others
who
are
really
still
trying
to
orient
and
figure
out.
How
do
we
approach
these
different
things?
A
So
the
first
one
is
really
about
accelerating
that
time
to
Market.
How
do
we
get
there
faster
and
how
can
we
make
sure
that
we
provide
that
ability
to
do
CI
CD
consistently
across
the
board,
so
that
our
app
devs
are
releasing
business
logic
and
being
able
to
really
augment
our
position
as
a
company
and
our
deliverables?
A
A
Basically,
everything
just
got
10
percent
more
expensive,
but
we're
not
making
more
money
based
off
of
it.
So
what
we
need
to
take
into
account,
then,
is
that
we
really
need
to
innovate
at
that
point,
because
we
can't
rely
on
what's
currently
in
place
to
maintain
that
structure
so
that
time
to
Market
becomes
really
important,
particularly
during
a
recession.
A
A
You
have
an
attacker
on
your
network,
so
then
it
becomes
a
question
of
how
do
we
enable
security
teams
to
set
policy
and
set
policy
as
code
so
that
the
rest
of
the
organization
can
run
within
those
confines
within
those
guidelines
and
make
sure
that
they
still
have
that
time
to
Market
and
then
last,
but
certainly,
not
least,
is
about
improved
operational
efficiency
by
improving
your
operational,
efficient
efficiency,
you
maintain
that
consistency
and
the
other
two
then
can
do
what
they
need
to
do.
You
have
the
security
you
have
the
time
to
Market.
A
A
A
If
I
give
them
238
choices.
I
promise
you
all
of
us
are
going
to
be
in
tears.
They're
going
to
be
in
tears,
I'm
going
to
be
in
tears,
I'm
gonna!
Guess
the
dog's
gonna
be
in
tears
because
they're
going
to
want
to
choose,
what's
shiny,
they're,
going
to
want
to
choose
what
tastes
good
they're
going
to
want
to
choose
what
they
think
is
logical
in
their
context.
A
A
So
if
we
want
to
translate
that
to
the
I.T
world,
how
are
we
enabling
Developers
making
that
simple
for
them
to
make
the
right
choices
so
that
they
can
support
the
business
value
and
by
supporting
that
business
value
and
making
that
more
simple,
we're
providing
more
room
for
Innovation,
but
room
for
Innovation?
That's
going
to
mean
something!
That's
going
to
hold
value
down
the
road.
A
I
need
things
to
be
simple
on
the
tooling
side
of
things,
I
also
want
things
to
be
more
simple:
on
the
people
side
of
things,
I
have
a
colleague
we
affectionately
call
him
Aquaman,
because
he
looks
like
Aquaman
and
one
of
the
things
he
always
asks.
His
customers
is.
If
you
had
one
wish
to
change
your
organization,
technical
people,
anything
what
would
that
be
now
for
me
without
a
doubt
without
a
question,
it
would
be
to
solve
the
people
problem,
because
that
is
the
most
complex
problem
that
exists
out
there.
A
Now,
if
we
put
in
con
Conway's
law
in
place,
what
we're
seeing
is
that
yeah
it
becomes
a
strategic
and
an
executive
decision.
So
how
can
we
create
some
of
these
multifunctional
teams?
What
we
know,
anthropologically
speaking,
is
that
humans
are
capable
within
any
particular
context
of
about
100
relationships.
A
If
I
look
at
some
of
these
organizations,
which
are
hundreds
of
thousands
of
people,
we
need
to
figure
out
how
we
can
boil
that
down,
so
that
we
have
the
correct
relationships
and
these
cross-functional
conversations,
because
we
also
know
that
cross-functional
teams
are
higher
performing
teams,
and
that
also
reduces
that
some
that
complexity
in
terms
of
people,
which
means
your
organization,
is
going
to
become
simpler,
Willow,
Run,
I'm
just
going
to
say
it.
You
can't
scale
a
unicorn.
A
Loop
speaks
for
itself
if
I'm
focused
on
the
outcome,
if
I'm
focused
on
complexity
getting
through
these
political
conversations
with
all
of
these
different
people
in
all
of
these
different
places,
because
Conway's
law
says
that
you
know
that
determines
my
architecture.
This
is
what
I
end
up
with
I
end
up
with
spaghetti.
I,
don't
want
spaghetti.
A
I
want
something
clean
I
want
an
iterative
loop
I
want
to
be
able
to
improve
my
architecture.
I
want
to
be
able
to
improve
my
development.
I
want
to
be
able
to
improve
my
organization,
but
if
I
don't
have
that
feedback
loop,
if
I
don't
have
the
velocity
of
that
feedback,
loop
I
can't
have
that
iteration
I
can't
have
that
feedback,
and
it
makes
it
a
lot
harder
to
drive
that
Innovation.
As
a
result,.
A
A
A
A
Now,
I
would
be
willing
to
say
if
all
of
you
go
back
and
you
look
at
your
change
management
and
the
things
surrounding
that
change
management.
Most
of
that
change
management
can
actually
probably
be
automated
industrialized
simplified,
and
you
will
only
unleash
a
lot
of
that
budget
in
terms
of
that
45
percent
and
if
we
look
at
Revenue
loss
just
based
off
of
misconfiguration,
because
people
are
doing
bespoke
things
within
organizations
and
not
following
following
best
practices.
A
A
What
we
soon
understand
is
that
there's
no
repeatability,
we
don't
know
who's
running
what
what
the
ecosystem
looks
like
security
people
aren't
happy
because
everything
is
bespoke.
They
can't
come
up
with
anything
to
follow
everything.
It's
not
scalable
operations,
people
very
similar,
they're,
saying
yeah
I
can't
follow
all
of
this.
This
is
all
different.
How
do
I
use
this
within
my
organization
and
networking
people
they're
all
like
I'm
supposed
to
do
all
of
this
by
hand.
This
is
all
different.
I
can't
manage
it.
How
do
I
do
this,
so
it's
not
sustainable.
A
So
then,
the
next
step,
and
once
we
realize
that
we
really
need
to
get
it
together,
create
something
a
little
bit
more
packaged
and
consistent.
It's
to
create
that
cloud
platform,
team
and
really
the
goal
here
is
to
create
that
underlying
functional
team
that
can
then
support
the
rest
of
the
team
so
that
when
you
have
your
application
and
development
teams,
you
have
an
embedding
of
somebody
from
that
cloud
team
to
help
you
make
that
connection.
A
A
So
basically,
what
we're
doing
is
simplifying
the
ecosystem
and
creating
a
broader
Funk
cross-functional
platform,
and
that
is
going
to
allow
us
to
collaborate
and
automate
drive
out
some
of
those
unicorns
and
really
just
help
us
migrate,
that
digital
and
physical,
so
that
they're
kind
of
more
Unified.
A
Now,
once
we've
done
that
with
public
cloud
and
what
we're
used
to
then
the
sky's
the
limit,
we
can
start
doing
this
with
whatever
we
want
to,
because
we
know
it
scales.
We
know
it
works.
We
know
it
starts
driving
out
some
of
these
bottlenecks
as
well,
and
you
have
that
collaboration
that
iterative
Loop,
the
Simplicity
is
there,
the
collaboration
is
there,
it
works.
A
A
A
How
are
you
simplifying
things
making
sure
that
it's
easy
to
pick
it
up
that
things
work
logically,
simplest
answer
is
usually
the
correct
answer.
Remember:
how
are
your
teams
communicating?
Do
you
have
an
open
line
of
communication
with
your
platform
team,
with
your
app
Dev
team,
with
your
SRE
team?
A
B
B
Sorry,
three
in
the
organizations
are
Italian,
and
so
we
have
Italian
times
I
will
we
are
going
to
We
were
supposed
to
have
a
15
minutes
break
to
to
get
from
downstairs
that
we
want
to
go
downstairs
and
want
to
go
up
stay
upstairs.
We
have
a
talk
happening
at
10,
past
10
in
downstairs,
and
we
are
going
to
have
other
eight
minutes
or
maybe
eight
minutes
break
and
then
we're
gonna
speed
with
the
guys
about
grpc.